Working Paper 2008-01

Land Regularization in
Tijuana, Mexico

Paavo Monkkonen

January 2008

UNIVERSITY OF CALIFORNIA
 Working Paper 2008-01

Land Regularization in Tijuana, Mexico

Paavo Monkkonen
Department of City and Regional Planning
University of California, Berkeley

Institute of Urban and Regional Development

University of California at Berkeley
I thank Tito Alegría and Gerardo Ordoñez for allowing me access to data
from their book, Legalizando la Ciudad.

2
 TABLE OF CONTENTS

Abstract ........................................................................................................5
I. Introduction............................................................................................7
II. Research on the Value and Demand for Land Titles
and Land Titling Programs ....................................................................9
III. Land Regularization in Mexico ...........................................................10
Land Regularization Tijuana, Baja California, Mexico.................11
Agency Organization .....................................................................12
Regularization Process...................................................................14
IV. Theory ..................................................................................................14
V. Empirical Analysis...............................................................................17
Results............................................................................................20
VI. Conclusion ...........................................................................................22
References..................................................................................................24

List of Tables and Figures

Table 1: Descriptive Information on Regularization Agencies

in Tijuana .................................................................................13
Figure 1: Miceli, Sirmans, and Kieyah Model of the Probability of
Land Registration.....................................................................16
Table 2: Contracting and Titling Rates by Agency................................18
Table 3: Summary Statistics of Regularization Contracting in
140 Neighborhoods..................................................................19
Table 4: Regression Results of Determinants of Percent Contracted ....21

3
4
 Land Regularization in Tijuana, Mexico
Paavo Monkkonen

Abstract
Land titling programs are increasingly encouraged by international
organizations as an essential component of urban policy in developing
countries. The clear definition of property rights is argued to be a sine qua
non of economic development. However, most academic research on land
titles has focused on the impacts of land titles and there is a dearth of
analysis of the demand for land titling and the structure of land titling
programs. A better understanding of land titling programs is essential for
the policy to succeed in improving the lives of people living in informally
developed neighborhoods.
This paper presents a model of demand for land regularization in
Tijuana, Baja California, Mexico. The model adapts previous work on
land registration in agricultural areas to regularization in urban areas,
incorporating insights on the different sources of value for urban land and
titles in urban areas, as well as the characteristics of the regularization
process in Mexico. An empirical test of the determinants of success in
land titling is conducted using administrative and spatial data from 140
irregularly developed neighborhoods where regularization agencies are
active. The prediction of previous models—that there should be more land
titling for land that is valuable—does not hold. The results demonstrate
inefficiency in the land titling system of Tijuana, inefficiency that is
possibly explained by governmental opportunism.

5
6
 Land Regularization in Tijuana, Mexico
Paavo Monkkonen

I. Introduction
While the importance of property rights to land for urban
development has been recognized for most of the 20th century, the work of
Peruvian economist Hernando De Soto (2000) galvanized the international
development community into placing a high priority on land titling
programs for urban policy in developing countries. Thus, understanding
the dynamics of land regularization and titling programs and the demand
for land titles should be increasingly important. Mexico provides an
exemplary case study, as a significant proportion of its urban areas were
developed irregularly and it has a long-standing regularization and titling
program. The present paper explores land regularization programs in the
city of Tijuana, Baja California, Mexico.
In order to understand the empirical evidence on the success of
land titling programs and the theoretical models of demand for land titles,
it is necessary to focus on the transaction between the government agency
and the resident. Using the lens of transaction-cost economics to analyze
land titling is beneficial because of its emphasis on the opportunism of
actors in the governance of contracts, contractual incompleteness and
flawed, second-best organization. Additionally, the four levels of social
analysis outlined by Williamson (2000) provide a useful framework for
the analysis of this transaction. Separating the formal rules governing the
transaction from the actual workings of the “play of the game” reveals
insight about contracting problems. For example, in the case of land
regularization in Tijuana, potential for governmental opportunism is
created by the variability in the fees charged and the minimal monitoring
of agencies by politicians.
To date, no model of demand for land regularization in urban areas
has been developed. Thus, in this paper, models of demand for land
registration of agricultural land in Brazil and Kenya, developed by Alston,
Libecap and Mueller (1999) and Miceli, Sirmans and Kieyah (2001),
respectively, are adapted to the demand for land regularization of urban
land in Tijuana, Baja California, Mexico.1 The two central predictions of
existing models—that the value of land titles increases with the value of

1
Regularization differs from registration in that is requires more technical work, such
as taking land surveys and creating parcel maps.

7
the land, and that the characteristics of residents are important—will be
determinants of the extent of land registration observed.
While the basic drivers of the demand for regularization are similar
to those for land registration, the regularization process is more
complicated and fees are generally levied on the lot holders. Additionally,
there are important differences in the sources of value for agricultural and
urban land. These differences lead to distinct predictions for the model
developed in this paper as compared to previous efforts. The main
contribution of the model is its consideration of programmatic features of
regularization, which is important because cost recovery is generally
emphasized in urban upgrading programs and results in significant
impacts on demand. Moreover, in the case of Tijuana, variability in
regularization fees creates opportunity for governmental opportunism. In
Mexico, lots in irregularly developed neighborhoods are regularized
through a contract between residents and a government agency. Residents
must agree to pay a fee for regularization proceedings, and these fees fund
the agency operations. Agency discretion over its funding source gives
incentive and opportunity to extract rents.
A test of a basic prediction of the model, using data from Tijuana,
produces results that contradict the basic theoretical predictions of
previous models and suggest the possibility of rent-seeking behavior on
the part of government agencies. There is a negative relationship between
the amount of land regularization in a neighborhood and indicators of the
value of the land in that neighborhood, controlling for a variety of other
factors. However, the result does not fully test the model because of a lack
of data and thus has some potential for bias. Nevertheless, it is a
provocative indicator of inefficiency in the regularization system and
perhaps opportunism on the part of government agencies.
The paper is organized as follows. The following section reviews
the literature on the demand for land titles. Section three presents an
overview of the land regularization process in Mexico and the land
regularization institutions in Tijuana. Section four reviews previous theory
and presents a model of the demand for land regularization for Tijuana.
Section five provides evidence on the success of land regularization
agencies and conducts a partial test of the model with data from 140
irregularly developed neighborhoods in Tijuana. The paper concludes with
a summary and directions for further research.

8
II. Research on the Value and Demand for Land Titles and
Land Titling Programs
In the international development community, property rights to
land have been an important topic for research during the past several
decades. Early academic work in this area focused on estimating the value
premium associated with legal title for plots of land. Studies in Indonesia
and Jakarta found high coefficients on a dummy variable for land title in
hedonic models of land value (Friedman et al., 1988; Dowall, 1991).
Recently, however, a more nuanced understanding has emerged of the
effects of different forms of property rights claims on the value of land.
For example, research in Ecuador demonstrated that informal property
claims also have a significant value premium (Lanjouw and Levy, 2002).
In the case of Ho Chi Minh city, where a full-scale formal property market
has yet to emerge after decades of government control, a combination of
informal claims and legal title actually has a higher premium than legal
title alone (Kim, 2004).
In addition to the value premium associated with land titles, recent
academic work has examined the impacts of holding a land title on the
various benefits ascribed to the formalization of property rights. In a
natural experiment of squatters in Buenos Aires, Galiani and Schargrodsky
(2007) found that land titles led to several hypothesized effects, such as
increased investments in housing, reduced household size, and improved
education of children. They did not find that titles led to increased credit
access or labor income, as did an analysis of the effects of the COFOPRI
land titling program in Peru, where Field (2003) found that titling
increased labor hours, especially work away from the home. The
discrepancies in this research suggest that the benefits of land titles depend
on context; for example, access to credit will not increase if banks do not
accept houses as collateral for loans or if people will not access formal
credit sources.
While the understanding of the benefits of titles deepens, research
on titling programs is still relatively new. The renowned COFOPRI
program in Peru assigned titles without charging lot holders; however, in
most countries, this will not be the nature of policy. Cost recovery
mandates will mean that some of the cost of the technical and
administrative work of titling, as well as some compensation for the
original landowner in the case of squatting, will be passed on to the
recipients of titles. In addition, in the case of illegal squatting rather than
illegal subdivisions, some compensation for the original landowner might
be charged to the residents before they are given titles, as occurs in
Mexico.

9
 Thus, an increasingly important component of research on titling
and regularization programs is a more complete understanding of the
demand for land titles. Recently, scholars have begun to focus their
attention on this question; however, the two existing models focus on
frontier and agricultural land exclusively. Alston, Libecap and Mueller
(1999) develop and test a framework for analyzing property rights on the
Brazilian Amazon frontier, and Miceli, Sirmans and Kieyah (2001)
develop and test a model of the probability of the registration of
agricultural land in Kenya. These models will be reviewed more
completely in section four.
Existing literature has not made a distinction between land titling
programs in agricultural areas and urban areas. This is surprising given
that land titles should be more important in urban areas than in agricultural
areas. Competition for land is greater in urban areas, and market
transactions that depend on the transferability and security of titles—sales,
loans, rentals and inheritance—are more common. Fixed capital
investment in urban areas is much larger than in rural areas. Investment in
land as a savings mechanism is also common, which makes security from
squatting more important. Additionally, the determinants of land values in
urban areas are different. In the model of Alston, Libecap and Mueller
(1999), the determining factor in land value was potential agricultural
productivity derived from natural characteristics, like the amount of
rainfall; in urban areas, however, the three principal determinants of land
value are its location relative to job centers and amenities, access to
services, and the characteristics of the neighborhood and neighbors.

III. Land Regularization in Mexico

Mexico has one of the most extensive and longest running land
regularization programs in the developing world. Massive rural-to-urban
migration after the Mexican revolution led to problems of squatting in
Mexico City as early as the 1920s. Land regularization programs began at
the federal level in the 1950s to regularize squatter settlements on state
land. A second regularization program was created for settlements on
ejido land in the 1970s. Ejido land is a communal land tenure system
established under Article 27 of the Mexican constitution of 1917. Under
this system, residents of the ejidos have agrarian rights to their land in
perpetuity, but the land cannot be sold, rented or mortgaged. The growth

10
of Mexican cities onto proximate ejidos has meant that they are frequently
converted to urban use illegally, which prompted a reform of Article 27.2
As in many countries, land regularization in Mexico is often a
politically driven process. Squatting is tolerated as a form of gathering
support of the urban poor. In the case of Mexico, through a clientilistic
system of favors for votes, the Partido Revolucionario Institucional (PRI)
was able to continually “win” elections and govern for the large part of the
20th century. One favor that was traded for votes was access to urban land
and services. Subsequently, regularization and titling programs were
instituted in these neighborhoods for continuing political support.
Apart from the political nature of the land regularization programs,
two further characteristics of land regularization in Mexico deserve
mention. While in many countries land regularization is important to
residents of informal settlements because it is a prerequisite for municipal
services, this is not true in the case of Mexico (Azuela and Duhau, 1998).
Because urban services are not dependent upon land regularization, the
demand for regularization is significantly affected; moreover, once
services are installed, they provide implicit tenure security from
government eviction. Secondly, slum clearance has never been a widely
applied policy in Mexico. Thus, demand for land regularization and titling
will be limited to the desire to use the land in market transactions.

Land Regularization in Tijuana, Baja California, Mexico.

Over one-half of the urbanized land in the city of Tijuana was developed
irregularly. There are over 160,000 lots currently registered with land
regularization agencies and only 270,000 households in the city (INEGI,
2000). Only about 60 percent of these lots have been contracted for
regularization; of those that have been contracted, only about 60 percent
have successfully completed titling procedures.3 There are three types of
land developed irregularly in Tijuana. First, land that was originally
owned by the federal, state or municipal government is often developed
irregularly, either squatted or converted to urban use as part of a program
for low-income housing. State-owned land that is slated to be developed
for public housing must go through a regularization process. Secondly,
ejido land that was urbanized must be expropriated and regularized.

2
In 1992, Article 27 was amended and ejido lands were allowed to be expropriated
from communal use and developed as private property; however, this development
modality has not been extensively utilized (Jones and Ward, 1998).
3
The data on irregular settlements used in this section and the rest of the paper are
taken from research conducted for Legalizando la Ciudad (Alegría and Ordoñez,
2005).

11
Finally, land owned privately is often developed irregularly, sometimes
with and sometimes without consent from the landowner.
The two principal models of irregular housing development in
Mexico and other parts of Latin America are the invasion of land by
squatters and the illegal subdivision and sale of lots by the landowner.
Land invasions can be on a large scale, organized groups of squatters who
settle a particular area in a coordinated fashion, or on a lot-by-lot basis. In
general, large-scale squatting, known in Mexico as invasión pirata, occurs
very infrequently these days. The form of squatting that continues to occur
regularly is the latter kind, known as invasión hormiga, and tends to occur
in areas where residential development is not permitted by law due to
steep slopes or other environmental conditions. The second mode of
irregular development, and the most common, is the illegal subdivision
and sale of land without regard for planning regulations or registry of
transaction. Settlements developed in this manner are not necessarily for
people with low-incomes, and are often indistinguishable from areas
developed following the legal requirements for residential development.

Agency Organization. There are currently six active land

regularization agencies in Tijuana with mostly separate spheres of
operation. They were created over the past fifty years by different levels of
government. Program administrators argue that several agencies are
needed to deal with the problem of irregular land development because of
the legal status of neighborhoods built on state, private or ejido land.
However, it seems likely that intergovernmental politics have played a
role in the creation of the agencies. Baja California often cites its distance
from the central government in Mexico City as a reason it is ignored.
Some say the central government’s neglect has led to relatively high
political independence in the city and state along with several innovations
in local governance (Ward and Rodríguez, 2000; Perló, 1999). Baja
California was the first Mexican state to elect a non-PRI governor, Ernesto
Ruffo, in 1989. Several mayors of Tijuana have also been from the
opposition party. Thus, the creation of the locally-controlled regularization
agencies, Fiduzet and PAR, was likely politically motivated and local
political administrations probably have had an involvement with federal
efforts in land regularization.
The first regularization agency in Tijuana, which is currently
known as CORETTE (Comisión de la Regularización de la Tenencia de
Tierra del Estado), was created by the federal government in 1957 to
regularize land in squatter settlements. Subsequently, separate agencies
were created for regularization of ejido land and land owned by different
levels of government. Recently, the municipal government created another

12
 regularization agency, PAR (Programa de Acciones de Regularización),
which will attempt to deal with illegal subdivisions through an amnesty
program for landowners who have already sold their lands without
following proper procedures.4 Descriptive information on the six agencies
is given in Table 1.

TABLE 1.
Descriptive Information on Regularization Agencies in Tijuana

Year Number Level of

Agency Created Type of Land of Lots Government Original Purpose Full Name

Main organization for

regularization of Comisión de la
invaded state land. Regularización
Private, State and
CORETTE 1957 58,000 Federal Has taken over de la Tenencia
Municipal
responsibility from de la Tierra del
other agencies in Estado
many neighborhoods.

Comisión de la
Regularization of
Regularización
CORETT 1978 Ejido and Federal 26,000 Federal illegal settlements on
de la Tenencia
ejido land.
de la Tierra

Took over
responsibilities from a
Inmobiliaria
federal agency that
INETT 1975 State 35,000 State Estatal Tijuana-
was created to
Tecate
develop low-income
housing.

Created by federal Promotora del

decree to satisfy Desarrollo
Produtsa 1983 Federal and State 13,000 State
demand for land for Urbano de
low-income housing. Tijuana, S.A.

Began as a state-level Fideicomiso

agency to regularize para el
neighborhoods on Desarrollo
Fiduzet 1992 Ejido 8,000 Municipal
ejido land and was Urbano de la
later transferred to the Zona Este de
municipal government. Tijuana

Recent municipal
effort to relax planning
regulations by giving Programa de
PAR 2002 Ejido and Private 24,000 Municipal amnesty to Acciones de
landowners who have Regularización
illegally subdivided
their land.

4
Because PAR was formed in 2002—the year that research for Legalizando la Ciudad
was conducted—data was not yet available on its domain. It is therefore not included
in the remainder of the analysis.

13
 Regularization process. The regularization process works in a
similar way for the six agencies and has four basic steps. First, an
agreement is signed with the original landowner to expropriate the land
and transfer it to the regularization agency. Second, surveys are conducted,
parcel maps are drawn, and planning approval is obtained. Third, contracts
for titling are written between the lot holders and the regularization agency.
Finally, the individual titles are paid for. The first step differs depending
on the original owner of the land, but once the land passes to a
regularization agency, the remainder of the process is the same.
In order to initiate regularization proceedings, public land has to be
unincorporated from government ownership and transferred to the
regularization agency. Similarly, ejido land must be expropriated from
agrarian use and transferred to CORETT. In the case of privately owned
land that was settled against the will of the owner, the owner can give the
land to the agency, which will then include the land costs in the fees for
regularization charged to the residents. After taking a commission (about
30 percent of the land value), the agency will pay the landowner fees
recovered. In the case of private land that was sold without permits, the
owner simply gives it to the agency. Once the land is transferred to the
agency, they conduct the standard administrative procedures for the
planning of subdivisions, surveying the land for parcel maps and a zoning
plan, and recording the land in the property registry.
The final two steps of the regularization process are transactions
between the regularization agency and the residents of the irregular lots.
First, the agency negotiates a contract with the lot holders for
regularization, under which they agree to pay fees for regularization
proceedings. Fees vary depending on the technical costs of regularization
for a given area and whether the land was squatted. Payments are
generally made in installments over a period of two to five years, and upon
full payment, the lot holders receive their titles. There is no foreclosure on
lots for which payments are not made, thus the incentive to pay stems
completely from the benefits of obtaining a title.

IV. Theory
Two studies serve as precedent for the model in this paper. Alston,
Libecap and Mueller (1996) developed a simple model of demand for land
titling in the Brazilian Amazon frontier. Their model is based on the
relationship between the value of a land title and the value of land. While
the empirical test performed in section five is similar to the Alston,
Libecap and Mueller model, the theory developed in the context of land
regularization in Tijuana is based on the work by Miceli, Sirmans and

14
Kieyah (2001). Their research is on the demand for land registration in
rural Kenya. Though the context is significantly different from land
regularization in urban Tijuana, the model requires only slight
modification.
Alston, Libecap and Mueller (1996) conceptualized the value of
agricultural land as determined by the production possibilities of the land,
which increase with proximity to a market. A title adds value to land by
enabling investment and exchange to expand production possibilities. This
added value declines as one moves away from a market center, as reduced
competition implies less need of state protection. At some distance from a
market, titled land will be of equal value to untitled land. In an empirical
test of their model, Alston, Libecap and Mueller demonstrate the above as
well as predictions that the cost of accessing a title increases at greater
distances from the administrative center. In addition, they test hypotheses
that the human and physical capital of claimants to land will affect the cost
of accessing land titles. Claimants with more education and wealth, for
example, have an easier time acquiring land titles.
Miceli, Sirmans and Kieyah (2001) formalize the relationship
between the cost of accessing land registration and the probability of
registering land by modeling the probability of land registration, p, as a
function of the cost of retaining a title, t. They assume the function p(t)
has a positive but diminishing marginal benefit of the input, thus p΄> 0 and
p΄΄< 0. Furthermore, p(0) ≤ 0 and p(t) < 1 for a finite t.
Miceli, Sirmans and Kieyah write the expected net present value of
land with the equation,
EV = pV – c(e,δ)t (1)
Where V is the dollar value of land and the costs of titling are a
function of education, e, and distance from the administrative center, δ.
Landowners are expected to maximize the expected value of their land. In
the Miceli, Sirmans and Kieyah model, the optimal title system for the
landowner is the pair ( p, t ) that maximizes utility, subject to the
constraint given by p(t). Utility increases to the northwest along the
indifference curves of the landowner, which are straight lines with slope
dp / dt = c / V > 0. This is shown graphically in Figure 1.5

5
A continuum of (p,t) pairs are assumed to exist and the optimum occurs at the
tangency of an indifference curve and the function p(t).

15
 FIGURE 1.
Miceli, Sirmans, and Kieyah Model of the Probability of Land Registration

p
p(t)
pB
B
pA
A

tA tB t

Varying the parameters V, e and δ gives comparative static results.

As V increases, the indifference curve becomes flatter, and the optimum
will shift from point A to B. Thus, owners of more valuable land demand
more protection even though this increases the cost of access. Assuming
that ce < 0 means that increases in education will lower the cost of
accessing land registration, also flattening the indifference curve from
point A to B. They also posit that areas closer to market and administrative
centers will have lower costs of accessing the system, thus cδ > 0. This
implies that as distance increases, demand for land registration decreases,
causing the owners’ indifference curve to steepen and the optimum to shift
from point B to A.
I use a model similar to Miceli, Sirmans and Kieyah; however, it
differs in two important ways. First, land regularization is different from
land registration. It adds a layer of complexity and potential for
opportunism on the part of the government. Because regularization
involves technical work for which fees are charged, and because the fees
charged to residents vary considerably, there is significant opportunity for
corruption. Evidence of corruption in land regularization programs in
Mexico has been documented in other sources (Varley, 1996). Secondly,
while education will continue to be an important component of system
access costs for residents of irregular settlements, income will also matter
considerably as it will determine the ability to pay for proceedings.
Therefore, using the same probability function, p(t), I write the expected
net present value of land with the equation,
EV = pV – c(e, Y, a, R(V), V*P)t (2)

16
 In the majority of the cases, the costs of obtaining a title are the
effort of negotiating a contract and an administrative fee, which in the
model are represented by the education, ce, and income, cY, of the lot
holder. I assume that ce < 0 and cY < 0, which means that increases in
education or income will flatten the indifference curves of lot holders and
shift the optimum from point A to B on Figure 1.
However, in this model, the pair ( p, t ) chosen reflects an
agreement between residents and regularization agencies on a payment
contract. The cost of regularization then will also depend on government
administrative costs, a, for which ca > 0. As costs of regularization
increase, indifference curves will steepen, shifting optimum from point B
to A. Additionally, government rent-seeking behavior will increase costs
of regularization as cR > 0. I propose that rent-seeking, R, is positively
correlated to the value of land, V. Because the value of a land title
increases as the value of land increases, opportunity for rent-seeking by
government officials will increase with the value of land.
While some predictions of the above model are clear, the
introduction of land value into the cost function creates ambiguity in the
predicted effects of land value on the probability of titling. It is clear that
increases in resident income and education will increase the probability of
titling and that it will decrease with the technical difficulty of titling. An
increase in land value will increase the probability of titling if the benefits
dominate the potential for governmental opportunism and the appraised
value of land, but whether this is the case is not clear from the model.
Finally, the costs will be greater for private land that was squatted.
Squatters must pay an appraised value of the land in addition to the
administrative fee. This extra charge is represented by cV*P, in addition to
the administrative fee. This cost only applies in certain cases, thus it is
multiplied by P, which is a dummy variable that equals 1 if the land was
squatted private land.

V. Empirical Analysis
At this time, existing data are insufficient to perform a complete
test of the above model; however, a partial test is possible. Unfortunately,
data describing residents of irregular neighborhoods are not readily
available because boundaries of census tracts do not correspond
sufficiently with the boundaries of irregular neighborhoods. Nevertheless,
the amount of land regularization activity in a neighborhood as a function
of indicators of land value and the technical difficulty of titling,
controlling for agency characteristics, is estimated. This corresponds to the
simple model of the value of land titles proposed by Alston, Libecap and

17
Mueller (1999). They theorized that the value of a title will increase with
the value of the land. However, in the case of Tijuana, a test of this theory
provides an interesting and counterintuitive result.
Before performing the test, one should note that the different
agencies exhibit wide variation in performance. Recall that the
regularization process has two final steps: contracts are negotiated
between an agency and residents on a lot-by-lot basis, and the residents
pay fees in exchange for titling. As Table 2 shows, some agencies, like
Produtsa, have contracted for titling with all or almost all the lots in their
jurisdiction, while others, notably CORETTE, have contracted with fewer
than half the lots on average. The number of lots that have actually been
paid off may be quite a bit less than the number contracted for titling—in
the case of INETT, only 35 percent.

TABLE 2.
Contracting and Titling Rates by Agency

Mean Percent Standard Mean Percent Standard

Agency Contracted Deviation Titled Deviation

CORETTE 45 .37 75 .25

CORETT 79 .16 66 .47

INETT 92 .12 35 .25

Produtsa 100 0 89 .09

Fiduzet 97 .04 86 .07

CORETTE and INETT have had the least success contracting and
receiving payment, respectively. It is likely that these differences arise
from the functions and institutional structure of the different agencies. For
example, it makes sense that INETT and Produtsa have a high rate of
contracting because they develop low-income housing through programs
of sites with services, and contracting occurs before residents inhabit the
piece of land. On the other hand, regularization agencies working in areas
where residents bought illegally subdivided land will have a harder time
contracting because they must negotiate payment for the land as well as an
administrative fee. People who purchased their land in an informal market
will resist paying again for it. In addition, the agency’s institutional
organization—i.e., their contracts with employees, institutional priorities,
and mandates from politicians—also influences its performance rates
significantly.

18
 Therefore, in order to test the previously elaborated theory,
controlling for agency fixed effects will be important. Unfortunately, we
do not have data on the exact administrative fee or appraised land value
charged for regularization. The available data show that it varies greatly.
There are two components, a regularization fee and a titling fee, both of
which depend on a variety of factors and differ between organizations,
neighborhoods and lots.
Data on 140 irregularly developed neighborhoods was obtained
from the five regularization agencies. In addition to the total number of
lots, the number contracted for titling, the number paid in full, the average
lot size and whether the land was originally owned by the government,
digital maps of the locations of the settlements were obtained. Using
digital mapping software (ArcMap and ArcView), the distance from the
city center was calculated along with the percent of land in a settlement
that has a slope of greater than 35 degrees. This is the legal maximum
slope for development in Tijuana. Table 3 presents summary statistics for
the sample.

TABLE 3.
Summary Statistics of Regularization Contracting in 140 Neighborhoods

Variable Mean St. Dev. Minimum Maximum

Percent contracted .52 .41 0 1

Log distance from city center 8.72 .61 6.84 9.91

Average lot size 442.5 718 76 5449

State land (dummy) .39 .49 0 1

Percent slope > 35° .19 .19 0 .76

Year regularization began 1987 8.39 1957 1999

The above variables were used to regress the percent of lots

contracted for regularization on a set of independent variables. Indicators
of land value and the technical difficulty of regularization, and controls for
the original owner of the land, the year regularization began, and agency
fixed effects were included. The indicators of land value included are the
natural log of distance from the city center and the average lot size.
Generally, half the variation in land value is explained by the distance
from the city center and the relationship is a negative exponential. Thus,
the expected relationship between the distance from city center and the
percent of lots contracted is positive; more valuable land should see more
contracting for regularization. Second, a dummy variable for whether the

19
land was originally owned by the government is included with the
hypothesis that it will be cheaper to regularize government land, as private
parties will not demand compensation. However, land owned by the
government will give more degrees of freedom to agencies negotiating a
contract, thus the relationship is not so clear. In addition to the proxies for
land value, a measure of the percent steeply sloped areas in a
neighborhood serves as a proxy for administrative difficulty in
regularization. It is expected that neighborhoods with lots of steeply
sloped areas will be more difficult to regularize. Finally, controls for
agency fixed effects are included.

Results. Table 4 presents the results of OLS and tobit models.

Tobit models are estimated because the dependent variable is a percent
constrained between values of 0 and 1. Tobit estimation corrects for bias
and inefficiency generated when using OLS on a constrained dependent
variable. While several independent variables perform as predicted, the
coefficients on indicators of land value have the opposite sign than the
simple model would predict. Neighborhoods closer to the city center with
smaller lots have a smaller percentage of lots contracted for titling. These
two variables are statistically significant to the .05 level in both OLS and
tobit models. This seems to indicate that government opportunism
dominates the incentives to contract for titling more valuable land.
Also of note is that the dummy variable for neighborhoods that
were developed on government land has a negative coefficient, indicating
that these have a lower percentage of lots contracted for titling. This is
also a counterintuitive result because, in these neighborhoods, residents
will generally have to pay less for their titles. Again, this suggests
government opportunism, as it is in these neighborhoods where agencies
have more flexibility in charging for regularization procedures.
The remaining variables have the predicted effects. Neighborhoods
with a large proportion of steeply sloped area have less contracting,
reflecting the increased administrative and technical difficulty in
regularizing land in areas where development is illegal. The year
regularization proceedings were initiated, included mainly as a control,
attains significance in the tobit model, indicating that more recently
developed neighborhoods have a slightly smaller percentage of lots
contracted for regularization. Finally, the dummy variables for
regularization agencies indicate that, ceteris paribus, CORETTE has a
significantly lower success rate in contracting for regularization.

20
 TABLE 4.
Regression Results of Determinants of Percent Contracted

OLS Model Tobit Model

Variable Coefficient T stat Coefficient T stat

Log distance from city

center .121** 1.97 .174** 2.29

Average lot size -.0001** -2.61 -.0001** -2.78

State land (dummy) -.135* -1.80 -.127 -1.41

Percent slope > 35° -.264* -1.74 .329* -1.79

Year regularization
began -.005 -1.38 -.008* -1.64

CORETT (dummy) .367** 2.57 .328* 1.92

Fiduzet (dummy) .316 1.59 .353* 1.40

INETT (dummy) .356** 3.66 .501** 3.97

Produtsa (dummy) .261** 2.62 .207* 1.73

2 2
Adjusted R .39 Pseudo R .31

N 140 N 140

F stat 10.78 Log likelihood -79.2

* indicates significance at the .1 level

** indicates significance at the .05 level

Two caveats must be considered in interpreting the above results.

There is a potential for omitted variable bias in the models due to the lack
of data on residents of irregular areas, and there is a potential for
ecological fallacy due to the aggregate nature of the data used. It could be
the case that the residents of the more valuable land are low-income and
lack education, which may be the dominant predictors of the extent of
regularization. However, this is not likely to be the case because, in
Tijuana, like much of Latin America, lower-income groups live on less
valuable land, further from city centers. The better-educated and higher-
income residents are more likely to live on greater-valued land closer to
the city center, which will actually increase the counterintuitive nature of
the results of the regression equations. The potential for ecological fallacy,
which is the threat that aggregating data at the neighborhood level
obscures relationships at the lot or household level, in the estimated
regressions is not large due to the relatively homogenous nature of

21
neighborhoods in Tijuana. However, this can only be overcome by using
household-level data.
A regression of the percentage of titles that have been paid for on
the same independent variables as the above regression was also
conducted. However, there is no statistical significance for the explanatory
variables, other than the agency dummy variables. Resident characteristics
are likely the principal determinants in whether the title is paid for or not,
though, again, it is surprising that the lots in more valuable areas are not
paid off at a higher rate.

VI. Conclusion
Understanding the nature of demand for land regularization and
titling in urban areas is increasingly important. Sustained interest in land
titling as an intervention in developing countries around the world stems
from the essential role that clearly-defined property rights play in the
functioning of market transactions. While research has demonstrated that
there is demand for titled land, the details of how regularization and titling
programs function has not received sufficient attention. Analysis through
the lens of New Institutional Economics reveals the importance of the
governance of the transaction between government agency and squatter or
resident of an illegally developed neighborhood.
Until now, analysis of the drivers and characteristics of demand for
land regularization has focused on agricultural land. The nature of demand
for land regularization and titling in urban areas is different from
agricultural areas due to the different sources of land value and the
increased importance of market transactions in urban areas.
Mexico is an ideal case for the study of regularization programs.
The system has existed for several decades and has characteristics that will
be problematic in most program designs—namely, the political nature of
land titling efforts and the potential for governmental opportunism in
charging title recipients. A model of demand for land titling procedures in
this context will differ from previously elaborated models not only
because of the characteristics of the Mexican titling system, but also
because of the nature of land value in urban areas. A partial test of the
model elaborated in the paper gives unexpected results; there is less land
titling in neighborhoods that are on more valuable land. This indicates an
inefficient system, possibly generated by governmental opportunism.
Though there is a potential for omitted variable bias, it is not likely, based
on evidence of the spatial distribution of income groups in Tijuana.
To truly unpack the dimensions of demand for land titles,
household-level data on the residents of informal neighborhoods is needed.

22
A multilevel model could be constructed that included neighborhood-level
data along with characteristics of households. This model would
disaggregate the effects of the various determinants of the variation in
titling between different neighborhoods. Testing such a model should be a
high priority for government agencies interested in pursuing land titling
programs as it could inform the reform of regularization proceedings.
Additionally, analysis of the governance of the regularization agencies
themselves and their relationship to politicians should be an important
component of future research.

23
References

Alegría, Tito and Gerardo Ordoñez. 2005. Legalizando la Ciudad. Tijuana,

Baja California: El Colegio de la Frontera Norte.
Alston, Lee J., Gary D. Libecap and Bernardo Mueller. 1999. Titles,
Conflict and Land Use. Ann Arbor, MI: University of Michigan
Press.
Azuela, Antonio and Emilio Duhau. 1998. “Tenure Regularization, Private
Property and Public Order in Mexico.” In Illegal Cities: Law and
Urban Change in Developing Countries, Edésio Fernandes and
Ann Varley, eds., pp. 157–171. London: Zed Books Ltd.
Demsetz, Harold. 1967. “Toward a Theory of Property Rights.” American
Economic Review 57(2):347–359.
De Soto, Hernando. 2000. The Mystery of Capital: Why Capitalism
Triumphs in the West and Fails Everywhere Else. London: Bantam
Press.
Dowall, David E. and Michael Leaf. 1991. “The Price of Land for
Housing in Jakarta.” Urban Studies 28(5):707–722.
Field, Erica. 2003. “Entitled to Work: Urban Property Rights and Labor
Supply in Peru.” Working Paper #223. Princeton, NJ: Research
Program in Development Studies, Princeton University.
Friedman, Joseph, Emmanuel Jimenez and Stephen K. Mayo. 1988. “The
Demand for Tenure Security in Developing Countries.” Journal of
Development Economics 29:185–198.
Galiani, Sebastian and Ernesto Schargrodsky. 2005. “Property Rights for
the Poor: Effects of Land Titling.” Working Paper No. 249.
Stanford, CA: Stanford Center for International Development.
INEGI. 2000. XII Censo de Población y Vivienda. Aguascalientes: INEGI.
Kim, Annette. 2004. “A Market without the ‘Right’ Property Rights.”
Economics of Transition 12(2):275–305.
Lanjouw, Jean and Phillip Levy. 2002. “Untitled: A Study of Formal and
Informal Property Rights in Urban Ecuador.” The Economic
Journal 112(482):986–1019.
Miceli, Thomas J., C.F. Sirmans and Joseph Kieyah. 2001. “The Demand
for Land Title Registration: Theory with Evidence from Kenya.”
American Law and Economics Association 3(2):275–287.

24
Perló, Manuel. 1999. “Mexicali: A Success Story of Property Tax
Reform.” Land Lines 11(5).
Varley, Ann. 1996. “Delivering the Goods: Solidarity, Land
Regularisation and Urban Services.” In Dismantling the Mexican
State, Rob Aitken, Nikki Craske, Gareth A. Jones and David
Stansfield, eds. London: MacMillan.
Ward, Peter M. and Victoria E. Rodríguez. 2000. New Federalism and
State Government in Mexico: Bringing the States Back In. U.S.–
Mexican Policy Report No. 9. Austin, TX: Lyndon B. Johnson
School of Public Affairs, University of Texas at Austin.
Williamson, Oliver. 2000. “The New Institutional Economics: Taking
Stock, Looking Ahead.” Journal of Economic Literature 38:595–
613.

25